App
The application surface: copilots, assistants and AI features people touch.
Read it as a grid. Down the side: the three assurance dimensions (App, Agent and Data). Across the top: the six lifecycle stages (Plan, Design, Build, Evaluate, Release, Run). Every cell is covered, because control isn’t something you apply once.
Wrapping the grid: the offerings at the top (Snapshot, Sign-Off, Managed) and the outcomes at the bottom: Visibility, Control, Evidence, Sign-off. The framework is the model behind the Snapshot, the Sign-Off Program, and Managed Assurance.
See where your AI sits, and what cannot be signed off yet.
Close the gaps until every system moves to approved.
Keep control in production as systems and risk change.
AI fails at the seams: a control applied at Build but never verified at Release, a dimension assured for the app but ignored for the agent behind it. A framework that covers every dimension at every stage is what closes those seams.
That’s the point of the full picture: not that any single cell is impressive, but that nothing is left uncovered between your AI ambition and your operational risk.